Ion channels give rise to many fundamental biological processes such as the generation of electrical signals in the central nervous system and the homeostatic regulation of cellular compartments. Channels have evolved to be opened and closed by a wide array of environmental cues making them ideally suited for their roles in such diverse biological phenomena. However, their core function lies in their ability to selectively facilitate the flow of ions across membranes. A few of their most critical properties are: the magnitude of the ionic flux (conductance), the ability to discriminate one ion over another (selectivity), and the current-voltage response (intrinsic I-V curve). A number of human diseases are linked to channel mutations that change their conduction properties. Several examples include: point mutations in the CFTR Cl- channel that reduce single channel conductance values and cause cystic fibrosis, a loss of K+ selectivity in TWIK-1 K+ channel which causes membrane depolarization and can lead to fatal cardiac arrhythmias, and sodium channel mutations that create cation leak currents that lead to periodic paralysis. Here, we will develop a computational tool for addressing these fundamental questions, by determining ion channel conduction properties directly from their atomic structures using all-atom molecular dynamics simulations. Due to the computational intensity of these calculations, past studies have failed to determine conduction properties under realistic conditions because they could not access timescales of physiological relevance. Using what is known as an enhanced sampling technique, we aim to overcome this shortcoming. We will first verify that our new approach rigorously reproduces previous current-voltage values reported for the pore domain of the KV1.2 potassium channel at non-physiological voltages. We then intend to show that the method is efficient enough to calculate currents at physiological membrane potentials, which has never been done before for a highly selective ion channel. Next, we will extend the method to directly calculate the K+-to-Na+ selectivity of the two-pore domain potassium channel, TRAAK, and elucidate how key structural elements of the protein influence the movement of different ion species through the channel. In accomplishing these goals, we will make experimentally testable hypotheses that will guide experiments and lead to a deep understanding of how TRAAK, and other narrow ion channels, operate at the molecular level. The flexibility of our method and its ease of use on widely available national computing platforms will enable other researchers to directly probe the permeation properties of any number of ion channels of known atomic structure.